AI Music Generator: Make Original Songs from a Text Prompt
VdoBloom’s AI music generator creates an original track from a short text description. Describe the genre, mood, tempo and instruments, optionally paste your own lyrics, and the AI composes a finished song with vocals or an instrumental that you can play and download as an MP3. It is built for creators who need background music for videos, a custom jingle, a birthday song or a quick demo of an idea.
Is it free? Every new account gets 10 free credits, enough for one ACE-Step track (6 credits), and the MP3 download works on the free plan. Each generation has a flat price: 6 credits on ACE-Step Base or ACE-Step Turbo, whatever length you choose, and 12 credits for a full song on MiniMax Music 2.6. The Lite pack’s 300 credits ($19.50 one-time, never expiring, or $15 a month) cover 25 MiniMax songs or 50 ACE-Step tracks.
Tracks render in the cloud, usually within a few minutes, and are saved to your My Creations library automatically.
How it works
- 1
Pick a music model
MiniMax Music 2.6 for complete songs with vocals and an instrumental option, ACE-Step Base for control over length and vocal language, or ACE-Step Turbo for faster results.
- 2
Describe the track
Write a style prompt with genre, mood, tempo and instruments, for example “upbeat synth-pop, 120 BPM, bright female vocals, punchy drums”.
- 3
Add lyrics or go instrumental
Paste your own lyrics for the AI to sing, or switch on Instrumental in MiniMax for music without vocals.
- 4
Set length and language (ACE-Step)
Choose a duration from 30 to 300 seconds and, optionally, the language the vocals should be sung in.
- 5
Generate and download
Click Generate. The finished track appears in the preview player with a Download MP3 button and is saved to My Creations.
Why use this tool
Vocals and lyrics
Both model families accept custom lyrics, so the vocals follow the words you wrote. Leave lyrics empty and the AI builds the song from your prompt.
Instrumental mode
MiniMax Music 2.6 has an Instrumental toggle for background tracks, beds for voiceovers and music for videos where speech must stay clear.
Tracks up to five minutes
ACE-Step lets you set any length from 30 seconds to 5 minutes, at the same flat 6 credits per track.
Flat price per track
No per-second billing: the cost is fixed per generation and shown before you click Generate.
Which AI music model should you use?
Start with MiniMax Music 2.6 when you want a complete, radio-style song with vocals, or a clean instrumental. Choose ACE-Step Base when you need a specific length, such as exactly 60 seconds for a reel or 3 minutes for a full track, or want to set the vocal language. ACE-Step Turbo uses the same controls and returns results faster, which suits trying several prompt ideas.
| Model | Best for | Lyrics | Instrumental toggle | Length control | Credits per track |
|---|---|---|---|---|---|
| MiniMax Music 2.6 | Full songs with vocals | Yes | Yes | Set by the model | 12 |
| ACE-Step Base | Detailed control, longer tracks | Yes | No (describe it in the prompt) | 30–300 seconds | 6 |
| ACE-Step Turbo | Faster drafts and variations | Yes | No (describe it in the prompt) | 30–300 seconds | 6 |
How much does AI music cost?
Each track has a fixed price, so a five-minute ACE-Step track costs the same as a 30-second one. One-time credit packs never expire; subscription credits reset each billing cycle. If a generation fails, the credits are refunded.
| What you make | Model | Credits | Tracks from Lite (300 credits) |
|---|---|---|---|
| Full song with vocals | MiniMax Music 2.6 | 12 | 25 |
| Instrumental background track | MiniMax Music 2.6 (Instrumental on) | 12 | 25 |
| 60-second reel soundtrack | ACE-Step Base or Turbo | 6 | 50 |
| 5-minute track | ACE-Step Base or Turbo | 6 | 50 |
How to write a music prompt that works
The style prompt steers the whole track. Specific, concrete words give better results than general ones like “good music”.
- Genre and era: “90s boom-bap hip hop”, “modern Afrobeats”, “lo-fi chillhop”.
- Mood and energy: “uplifting”, “melancholic”, “tense and cinematic”.
- Tempo: give a BPM, such as “85 BPM” for lo-fi or “128 BPM” for house.
- Instruments: “warm Rhodes piano, upright bass, brushed drums”.
- Vocals: “soft male vocals”, “powerful female vocals”, or “no vocals”.
- Lyrics: label parts with [Verse] and [Chorus] and keep lines short and singable.
Read more: AI lo-fi music ideas · AI rap song generator
Prompt examples to start from
Copy one of these, then change the details to fit your project. Pick the model in the third column for the closest match.
| Use | Example style prompt | Model |
|---|---|---|
| YouTube background | Calm lo-fi hip hop, 80 BPM, soft piano, vinyl crackle, no vocals | MiniMax Music 2.6 (Instrumental on) |
| Birthday song | Cheerful acoustic pop, ukulele and claps, warm female vocals, singalong chorus | MiniMax Music 2.6 with your lyrics |
| Reel soundtrack | Energetic EDM build and drop, 128 BPM, bright synths | ACE-Step Base, 30–60 seconds |
| Brand jingle | Upbeat funk, slap bass, brass stabs, short catchy hook | ACE-Step Turbo, 30 seconds |
| Cinematic intro | Epic orchestral, slow build, taiko drums, choir swell | ACE-Step Base, 90 seconds |
| Regional song | Romantic Bollywood ballad, tabla and strings, male vocals | ACE-Step Base, vocal language Hindi |
How AI music generation works
Music models learn the patterns of melody, harmony, rhythm and vocal phrasing from large amounts of audio. When you send a style prompt, the model composes a new piece that matches the description, arranging instruments and, if you supplied lyrics, fitting the words to a melody and singing them. Each generation is new, so running the same prompt twice gives two different takes.
That randomness is useful: if the first version is close but not right, generate again or adjust one detail in the prompt, such as the tempo or the lead instrument, rather than rewriting everything.
What creators make with it
Most tracks end up under a video or as a standalone song. Popular uses include:
- Background music for YouTube videos, Reels and TikToks
- Custom songs for birthdays, weddings and gifts, with the person’s name in the lyrics
- Jingles and short brand themes for ads
- Music beds under podcast intros and voiceovers
- Demos that turn written lyrics into a sung draft
- A soundtrack to turn into a music video
Read more: How to add music to a video with AI · Turn a song into a music video
Limits to know
A few things the music generator does not do on this page:
- Output is a single mixed MP3. To split a song into vocals, drums, bass and other stems, use the Vocal Remover and Stem Splitter.
- You cannot upload a song to restyle or extend here; that is what AI Song Cover does, for songs you own.
- MiniMax Music 2.6 has no duration setting; use ACE-Step when you need an exact length.
- Do not ask for a living artist’s voice or a copyrighted song’s lyrics; describe the style instead.
Frequently asked questions
Is the AI music generator free?
You can try it free. New accounts get 10 credits, enough for one ACE-Step track at 6 credits, and the MP3 download is available on the free plan. MiniMax Music 2.6 songs cost 12 credits. The Lite pack’s 300 credits ($19.50 one-time) cover 25 MiniMax songs or 50 ACE-Step tracks.
Can the AI sing my own lyrics?
Yes. MiniMax Music 2.6, ACE-Step Base and ACE-Step Turbo all accept custom lyrics next to the style prompt, and the vocals follow your words. Mark sections with [Verse] and [Chorus] to shape the structure.
Can I make instrumental music with no vocals?
Yes. Switch on the Instrumental toggle in MiniMax Music 2.6 to get a track without vocals. With ACE-Step, leave the lyrics empty and describe the track as instrumental in the prompt.
How long can an AI-generated song be?
ACE-Step Base and Turbo let you choose any length from 30 to 300 seconds, so up to five minutes, at a flat 6 credits. MiniMax Music 2.6 decides the length of the full song itself.
Can I use AI-generated music on YouTube or in commercial projects?
Under VdoBloom’s terms of service, you retain ownership of the content you generate with the service. Check that your prompt and lyrics do not copy someone else’s song or imitate a real artist, and read each platform’s rules for AI-generated music before monetizing.
Which AI models does VdoBloom use for music?
Three models: MiniMax Music 2.6 for full songs with vocals and an instrumental option (12 credits), ACE-Step Base for detailed control over length and vocal language (6 credits), and ACE-Step Turbo for faster generation (6 credits).
Can I make songs in other languages?
Yes. Write the lyrics in the language you want, and with ACE-Step you can also type the vocal language, such as Spanish, Hindi or Japanese, in the Vocal Language field.
How long does it take to generate a song?
Most tracks finish within a few minutes. ACE-Step Turbo is the fastest option. You can leave the page while it renders, because every track is saved to My Creations.
What file format do I get?
Tracks download as MP3 files from the preview player, ready for any video editor, DAW or VdoBloom’s music video tools.
Can I turn the song into a music video?
Yes. Use the AI Music Video Maker to cut visuals to the beat of your track, or the AI Music Video Generator to turn a photo into a performance of the song.